steelman-then-break
Pressure-test a design or tech choice in two forced passes — strongest case for, then strongest case against, then a calibrated verdict that's willing to say 'proceed'.
Install
$ curl -fsSL https://raw.githubusercontent.com/Osipchuk/agent-skills/main/install.sh | sh -s -- steelman-then-breakOne line, works with any agent — installs just this skill into ~/.claude/skills/. Needs only curl; it fetches uv and the askill CLI if they're missing.
$ /plugin marketplace add Osipchuk/agent-skills
$ /plugin install skills@askillTwo slash commands inside Claude Code — installs the whole library as one plugin and lets Claude pick the right skill by context. No permission prompts, no remote-code-execution dance.
Installs skills/steelman-then-break/SKILL.md into
~/.claude/skills/steelman-then-break/; Claude reads it on the next
session start. The plugin tab pulls the whole library instead.
When it fires
Fires on "review this proposal", "should we go with X", "steelman X", "red-team this design", or when choosing between an architecture, library, or framework. It forces the second pass that design discussions skip, and pulls prior failures from references/scars.md so objections argue from the team's own evidence.
name: steelman-then-break
description: Pressure-test a design proposal or technology choice in two separated passes: first the strongest possible case FOR it (steelman — argue the best version, even stronger than the author put it), then a rigorous case AGAINST it (red-team — name the anti-pattern, lead with what breaks first, cite the team's prior scars), then finish with a calibrated verdict. Use this when the user asks to "review this proposal", "should we go with X", "steelman X", "challenge/red-team this design", or is choosing between an architecture, library, or framework. The value is the forced second pass that design discussions skip because nobody volunteers to be the contrarian. Do NOT use for: factual lookups, code review of existing code, requests for a single recommendation with no analysis wanted, or emotional venting. Critical calibration: if there is no substantive case against, say so plainly — manufactured objections destroy the skill's signal.
What it does
- — Forced separation: the case-for has no caveats, the case-against leads with what breaks first — no hedging mush.
- — Cites the team's own scars (references/scars.md), so the red-team argues from evidence not generic risk.
- — Calibrated verdict: willing to conclude 'the steelman holds, proceed' instead of reflexively finding fault.
- — Built-in guard against the failure mode — manufactured objections destroy the signal, so it says when there's no real case against.